Institute of Transportation, MOTC

     Lane-changing is regarded as a risky driving task. As the process will closely interact with the surrounding vehicles, it will increase the collision risks. Yet little research has explored lane-changing behavior in different driving situations, particularly in urban roads, which has a challenging driving environment. This study focus on lane-changing decision-making and collision risks. Two analyses are made to identidy the frequently changing lane situation and the high-risk situation respectively. Finally, find out the both frequently and high-risk lane changing situation through the quadrant matrix diagram. This research uses millimeter-wave radar installed at urban intersections to collect microscopic vehicle trajectory data. First, the lane change event is extracted by observing the real-world driving behavior, and when the vehicle starts to change lanes, the driving dynamics of the vehicle ahead and the vehicle behind in the target lane are classified into 16 driving situations, and then the stopping distance index (SDI) is used to measure whether a collision risk occurs. The results of frequently situations analysis found that the most likely scenarios for changing lanes are as follows: (1)A safe driving distance between vehicles is not maintained.(2)When the front and rear vehicles are close to each other. From the high-risk situations analysis, it is found that the high-risk scenarios for changing lanes are: (1)The vehicle does not maintain a safe driving distance between vehicles.(2)When the front and rear vehicles are approaching each other although a safe distance is maintained. Based on the above analysis results, it can be further found that the both frequently and high-risk situations where changing lanes are likely to occur are: (1)Not maintaining a safe driving distance. (2)Although a safe distance between vehicles is maintained, the situation where the oncoming vehicle is approaching the vehicle is not noticed. In addition, through analysis, it is also found that the above-mentioned scenarios are more likely to occur at intersections, changing lanes to the left, and high traffic volume.